Hacker Newsnew | past | comments | ask | show | jobs | submit | ronsor's commentslogin

Shorter links.

Or for historical context, this is all Twitter's fault since its character limit included links too.


Before twitter, it was forum signatures with a character limit :’)

Worth noting it doesn't anymore, all links are now worth the same by going through Twitter's own link shortener.

> VARA’s developer released a new version 4.9.1 which had a logic bomb that completely breaks compatibility with all prior versions starting on 1 Jan 2027.

Is there more coverage? That sounds downright evil.


Nothing official from the VARA developer afaik.

I participated in some discussion on a private discord where the claim was made that the developer did this. I reversed the binary with an LLM and confirmed it myself. I’m aware of at least one ham radio YouTuber who has posted a video about this.

I presume it will get more coverage as more people learn about the new release and the compatibility break.


Official VARA site: https://varamodem.com/

If you can find anything like release notes you are doing better than I did.

A web search for

"VARA version 4.9.1 incompatibility"

brings up as much coverage as you are apt to find.


User Mode Linux does work, but it can have significant performance issues and isn't actually very portable.

Yes, I've tried using it before.


He literally acknowledged that:

> WireGuard by itself is not a stealth VPN, but the things that make a VPN "stealthy" are all things you'd layer on top of a WireGuard design, not thread through the middle of it. A datagram-based protocol can run under any obscured transport.


Reminder:

https://fixupx.com/v01dpr1mr0s3/status/2119993017717358913

> Anthropic injects a hidden <ip_reminder> in Fable 5.2's turns via API;

> Fable gets obsessed over it and Anthropic blocks the request for "reverse engineering".


It works perfectly well as long as you realize some machines are simply broken.

For example, ones in China simply never can pull images. I don't know how they got approved!


> is ability to combine tool calls a useful thing to measure, i.e. should it confer an advantage in the benchmark?

Yes, models are generally rewarded for this during RL. Not combining tool calls is an objective deficiency, even if not a major one.


Yes, two ways failing to combine tool calls is a real deficiency for the LLM's user: it costs money and can get you rate limited.

Every turn forces you to pay $ for cached inputs and possibly gets you rate limited on TPM. At 500k context and $0.26/M-tok cache read price, every time it combines two tool calls instead of taking separate turns would save you $0.13. That could add up.

Maybe the provider sets a rate limit of 2 million TPM. That's just 4 turns in 1 minute before you start getting 429s. Cutting them in half by combining calls can save that.

All else equal, we definitely want models combining tool calls!


It also makes models completely ungovernable and unfollowable.

Asking for user permission before an action is pointless. it's not going to use a single `write`, it's going to run a 25 line perl script without even newlines.

At this point just completely remove permissions, and leave only one tool, a rust-like scripting language.

But this also drives usage up, because you rely less on yourself, and more on the increasingly black-box that is the subagent-of-subagent that you are not even going to check the result of.


It depends on what you actually want to ask permissions for. We're still doing trivial script execution, but we could go further. Capabilities like "do you allow giving access to JIRA read only token" would be a massive improvement. But there are other paths too.

Polymarket didn't serve the US last I checked, unless they've changed recently.

Polymarket does. In the US they are a registered CFTC exchange for trading event contracts. That gives them some degree of protection from state actors. It also increases the odds that if you make money you'll actually be able to withdraw it unlike many of these markets. They also operate in actual dollars not make up crypto tokens.

Note that Polymarket's non-US business is crypto blockchain whatever so ... good luck with that. The ability to get your money out could disappear at any time and as far as I know there's no recourse if they just decide to screw you over. Not that CFTC is super effective but it is something.


> In the US they are a registered CFTC exchange for trading event contracts. That gives them some degree of protection from state actors

So is Kalshi. The entire question is how the CFTC made up the authority to regulate gambling out of statute that only ever mentions exempted commodities.


They do nothing to block US people from signing up to polymarket.com, the block is a farce

Ah, makes sense. I've never bothered looking into using either so I wasn't aware of that.

They do: polymarket.us

"AI llms" is an awkward phrase I've never seen before.

About never, because ADS-B broadcasting is required and it's trivial to receive with a cheap (<$30) SDR.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: